Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add MauricioPerera/KDD --skill delegar-ollamagit clone --depth 1 https://github.com/MauricioPerera/KDDWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/mauricioperera/kdd/delegar-ollama)<a href="https://agentmods.dev/skills/mauricioperera/kdd/delegar-ollama"><img src="https://agentmods.dev/badge/skills/mauricioperera/kdd/delegar-ollama/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/mauricioperera/kdd/delegar-ollama"><img src="https://agentmods.dev/badge/skills/mauricioperera/kdd/delegar-ollama.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00000 | $0.02604 |
| Opus 5 | $0.00000 | $0.01302 |
| Sonnet 5 | $0.00000 | $0.00521 |
| Haiku 4.5 | $0.00000 | $0.00260 |
Grade A, and why
delegar-ollama scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 74 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Delegar a un modelo vía Ollama (headless) — mecánica agnóstica al modelo
Patrón de lanzamiento reutilizable, independiente del modelo (<MODEL>: glm-5.2:cloud, kimi-k2.7-code:cloud, etc.) y del repo (<REPO>: cwd salvo que se indique otro). Esta skill NO sabe de CCDD, contratos ni de qué tarea estás delegando — solo de cómo hablarle a un modelo por Ollama sin que se cuelgue o devuelva vacío.
Comando base
ollama launch claude --model <MODEL> -y -- --strict-mcp-config --mcp-config '{"mcpServers":{}}' -p "$(cat <prompt.txt>)" --dangerously-skip-permissions < /dev/null > <log.txt> 2>&1
< /dev/nullOBLIGATORIO. Sin él,claudeespera stdin y el lanzamiento sale VACÍO (exit 0, sin trabajo).--strict-mcp-config --mcp-config ...OBLIGATORIO (verificado 2026-07-03/04): sin él, cada lanzamiento hereda y levanta TODA la flota MCP global del usuario (n8n, github, chrome, computer-use...) — decenas de procesos que colgaron la app anfitriona dos veces (evento Windows 1002 "dejó de interactuar"). Regla: si la tarea NO usa el gate CCDD →'{"mcpServers":{}}'; si lo usa → un JSON con SOLO la entradaccdd-complexitycopiada de~/.claude.json.- Prompt en ARCHIVO +
$(cat ...), nunca inline: evita problemas de comillas/multilínea. - Permisos:
--dangerously-skip-permissionses lo más simple (repo desechable). Alternativa más contenida:--permission-mode acceptEdits --allowedTools "Bash,mcp__ccdd-complexity__*"— ojo queacceptEditsSOLO auto-aprueba Edit/Write; en headless no hay quién apruebe Bash/MCP, así que sin--allowedToolsexplícito el modelo reescribe archivos pero NO puede correr tests/gate y cae a "autoevaluación estática" (PASS falso). Usá--allowedToolscon exactamente lo que la tarea necesita. - El modelo hereda el cwd desde el que lanzás —
cd <REPO>antes, o pasalo en el prompt. - Tareas largas:
run_in_background: true, leé el log/.outputal llegar la notificación de fin.
Variante /goal (verificado 2026-07-02, claude ≥ 2.1.139)
Para que el modelo siga trabajando SOLO hasta cumplir una condición (un evaluador independiente juzga tras cada turno y lo devuelve a trabajar con feedback si no cumplió):
ollama launch claude --model <MODEL> -y -- -p "/goal <condición>" --dangerously-skip-permissions < /dev/null > <log.txt> 2>&1
- La condición: estado final + chequeo demostrable en la conversación + restricciones + tope OBLIGATORIO ("o parar tras N turnos", máx 4.000 chars) — sin tope no hay límite de gasto.
- El tope de turnos NO protege bucles DENTRO de un turno. El evaluador corre al TERMINAR cada turno; un modelo que loopea llamando tools nunca termina el turno y el evaluador jamás se ejecuta (caso real: 992 reintentos de una misma tool, 67 min). El timeout del task en background TAMPOCO lo mata. Único guardián real: monitorear el transcript en disco (¿crece sin avanzar? → matar el proceso).
- El evaluador NO ejecuta comandos: solo juzga lo que el modelo mostró — la condición debe exigir salida REAL pegada. Veredicto queda como
goal_statusen el transcript. - Reemplaza la re-delegación manual por FAIL dentro de una misma invocación (reintentos automáticos con la razón del evaluador como feedback). Costo del evaluador: despreciable (~1,5k tokens/veredicto).
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 74 lines · 0 tokens per session scan A 274aeeff7020
delegar-ollama is a skill published in the GitHub repository MauricioPerera/KDD (8 stars, last pushed 3d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 2,604 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
crit-cli
Use when an agent needs to author or reply to crit inline comments programmatically (including multi-agent workflows commenting on shared code/plans/docs/proposals), publish or unpublish a crit review with crit share, sync a crit review to or from a GitHub PR or GitLab MR, or read/interpret a crit review JSON file.…
crit
Review code changes, a plan, a live page (running dev server), or a local HTML file with Crit inline comments and structured human feedback. Use only when the user explicitly invokes /crit or directly asks to use Crit; a generic review request does not count.
crit-story
Author a crit story and continue the interactive review loop only when the user explicitly invokes crit-story or directly asks you to generate a crit story. Do not infer this skill from generic review, PR, or diff-review requests.
trio-omnigent
Run the Cursor-backed Omnigent Trio loop from the current Claude/Codex UI session when the user explicitly says “Trio Omnigent”, “Omnigent Trio”, or invokes /trio-omnigent. Do not use for an ordinary native Trio request.
trio-bridge
Suggest background Trio work by scanning a project and writing deduplicated proposal cards into proposals/.
research-a-evolve
Provides guidance for automatically evolving and optimizing AI agents across any domain using LLM-driven evolution algorithms. Use when building self-improving agents, optimizing agent prompts and ...